Datareader doesn't download the desired samples as instructed

[yongsua1989]

Hi, I would like to download only the stock price data for 3 months from Google Finance, but it keeps providing me data for one year. How do I slice the data I want or re-download the data I want?

from pandas_datareader import data
from datetime import datetime

cname = "KLSE:AXIATA" # Yahoo:6888.KL #Google: KLSE:AXIATA

wname = "google"

start= datetime(2017, 7, 15)

end=datetime(2017, 9, 15)

price = data.DataReader(cname, wname, start=start, end=end)["Close"].pct_change()

[yongsua1989]

Anyway to slice the sample data? Is it possible instead of downloading whole year of data?

[rsvp]

The temporary problem with specifying dates is diagnosed here: #394 (comment)

Getting more data than you need should not be a major problem with pandas, e.g. use

 DataFrame.tail(n=64)

to approximate 3 months of daily data.

[i-frost]

I am doing this to filter for specific dates

df = df_reader.to_frame(filter_observations=False)

df[(df.index.get_level_values('Date')>='2017-09-21') & (df.index.get_level_values('Date')<='2017-09-22')]

[rsvp]

@i-frost as of v0.20.1 Panel is deprecated: https://pandas.pydata.org/pandas-docs/stable/whatsnew.html#whatsnew-0200-api-breaking-deprecate-panel and thus ".to_frame".

For contiguous dates, the easier expression would be: df['2017-09-21':'2017-09-22'] which relies on index slicing.

[i-frost]

Hey @rsvp, I am getting a Panel back from the reader call should it be a MultiIndex instead?

pd.version
'0.20.3'

df_reader = web.DataReader(['III'], 'google', (2017, 9, 23), (2017, 9, 23))
df_reader
<class 'pandas.core.panel.Panel'>

[HJ1nitj]

Hi,
I am getting this 'ImmediateDeprecationError(DEP_ERROR_MSG.format('Yahoo Daily'))' while fetching the financial data from yahoo.

data=web.DataReader(stocks, data_source='yahoo', start=start_date, end=end_date)['Adj Close']

please help me with this issue.

[bashtage]

This is not fixable. Yahoo is totally broken until someone figures out a way to rewrite it from scratch using a different method.

[HJ1nitj]

So can you please suggest some alternative?

[bashtage]

You can manually download data from Yahoo in the browser. Or you could see if the data you need is available from one of the other sources (e.g., Morningstar, IEX, etc.). AFAIK there is no access to Yahoo from Python that works now.

[HJ1nitj]

thanks sir...